The Weekly Newsletter of MIT Linguistics

MIT Linguistics Colloquium - Henk van Riemsdijk - April 18th

Friday, Apr. 18, 3:30 PM

Henk van Riemsdijk
Tilburg University

“Parameterizing Laws of Nature: Some Thoughts on Identity Avoidance”

In earlier work (Van Riemsdijk, 1989), I had argued that headed relatives in Swiss German do not involve wh-movement. Instead there is an invariable relative complementizer wo and regular pronominals are used as resumptive pronouns. These pronouns tend to be clitic-like, and as such they can be adjoined to the C°-position. And if that C°-position is the one adjacent to the head of the relative clause, then a configuration of local licensing for deletion is obtained. One major question arising from this state of affairs is that deletion is absolutely obligatory in the sense that clitic movement, normally optional, must apply here to feed deletion. This was handled in Van Riemsdijk (1989) by means of a global version of the Avoid Pronoun Principle.

Two other major problems had been largely ignored in my earlier work. First, the resumptive pronoun is not obligatory as such. In fact, part-whole relations will suffice to establish a semantic connection between the head of the relative clause and the relative clause itself. In this sense, Swiss relatives are quite similar to English such that relatives like ‘a triangle such that the sum of the squares of the two shorter sides equals the square of the longest side’. The question then arises whether there is any reason (presumably syntactic and not semantic) to assume that some covert correlative element is involved in such cases.

Second, there is one other situation (in addition to the deleted clitic) in which a gap is found. This is when the correlative element is a locative. The overt locative wh-word is wo, that is, it is identical to the relative complementizer. The relation between the locative gap and the head of the relative clause is characterized by the usual movement diagnostics. We must conclude therefore that wh-movement is involved in this case. Predetermining the choice of the strategy (wh-movement, resumptive pronoun, mere semantic aboutness) in the derivation of Swiss German headed relative clauses thereby becomes a serious problem. What seems to be going on, quite patently, is that wo when moved to Spec,CP, can then be deleted in a process of haplology, a kind of OCP-effect in syntax, cf. Van Riemsdijk (1998). Thereby we have established a situation rather reminiscent of the generalization concerning clitic resumptive pronouns: wo must move in order to be deleted.

I will also argue that the aboutness cases that apparently lack a correlative element altogether actually involve a locative (wo) expletive adjunct that also moves only to be deleted. It appears, then, that annihilation of the correlative element is the true generalization underlying the analysis of headed relatives in Swiss German. The major principle forcing deletion is the Doubly Filled COMP Filter, which can also be interpreted as a principle that avoids (relative) identity. I will conclude with a discussion of the possible status of such a general principle of ‘identity avoidance’ as a general principle of design that is at work in (among others) grammar.